The problem was cavalry, not commerce
TeaHQ's own reading of well-attested general knowledge, with no single source behind it.
The Song faced a military situation its predecessors had partly avoided: powerful mounted neighbours to the north and west, and no reliable domestic supply of the kind of horses cavalry needs. Horses of that quality come from grassland, and the grassland was under the control of the states the horses were needed against. A settled agrarian empire with a large fiscal capacity therefore had to buy them, and had to buy them with something the grassland peoples wanted and could not make. Tea was the answer, for reasons that are entirely practical: the peoples of the high plateau and the steppe had adopted it, it does not grow in their territory, it keeps and travels in compressed form, and China could produce a great deal of it. The tea-and-horse exchange is best understood not as a trade route with a romantic name but as a procurement programme with a defence objective, run by officials with quotas.
The instrument was a monopoly, and it had to be
TeaHQ's own reading of well-attested general knowledge, with no single source behind it.
For the exchange to work, the state had to control the supply. If private merchants could sell tea across the frontier at will, the price of tea in horses would be set by the market and the state would lose its leverage; worse, the neighbours would obtain tea without giving up horses. So the tea of the producing districts nearest the frontier — Sichuan above all — was brought under state purchase and its private export prohibited. The standard account dates the formal apparatus to the 1070s, under the reform administration, with a dedicated Tea and Horse Agency established to buy Sichuan tea, move it west, and exchange it at designated markets. TeaHQ gives that date as the conventional one. What is clear regardless of the exact year is the logic: the monopoly was not a revenue device that happened to involve horses, it was a horse-procurement device that required a monopoly to function.
The exchange rate was policy, not price
TeaHQ's own reading of well-attested general knowledge, with no single source behind it.
The number of catties of tea a horse was worth was set administratively, adjusted repeatedly, and differed between markets and grades of horse. That is the most revealing feature of the whole system. A rate set by an agency is an instrument: raise the tea price of a horse and you extract more from the sellers but risk them going elsewhere or refusing; lower it and you spend more tea than you need. Officials argued about it in exactly those terms. The consequence is that Chinese frontier tea policy contains, several centuries early, an explicit governmental theory of terms of trade — how much of a manufactured good to give for a strategic import, and what happens to the counterparty’s willingness to supply when the terms move. Anyone who has followed a modern commodity-policy argument will recognise the shape, and the reason it is legible at all is that the state wrote its reasoning down.
What it did to Sichuan
TeaHQ's own reading of well-attested general knowledge, with no single source behind it.
Being the designated supplier to a state monopoly is a mixed inheritance for a producing region. It assures demand, at an administered price, for a defined specification — which in practice meant coarse leaf, compressed hard, made to survive a long overland journey by porter and pack animal into the mountains. It removes the incentive to make anything else with that leaf, because the buyer is not paying for refinement. And it ties a large share of a region’s agricultural economy to a policy decision made a long way away. Ya’an and the western Sichuan tea districts carry that history: a long specialisation in border-trade tea, a manufacturing tradition built around compression and long ageing in transit, and a product whose modern descendants are still made to a frontier specification rather than to a connoisseur’s one. The Yunnan branch of the same trade left a comparable mark, and is a large part of the prehistory of pu-erh.
What it left behind
TeaHQ's own reading of well-attested general knowledge, with no single source behind it.
Three things outlived the bureau. A permanent frontier tea manufacture, still producing compressed dark tea for the plateau markets, with grading conventions that come from a procurement specification rather than from a tasting. A political relationship in which tea supply is leverage — one that reappears in the Qing period, and again when British India tried to push its own tea into the Tibetan market against a supply chain that had been running for centuries. And a route network that a much later tourism vocabulary would rename the Tea Horse Road, a coinage of the modern period rather than a historical name for the system, discussed further in `tradition-tea-horse-road`. The bureau itself is the part that usually gets lost, which is unfortunate, because an agency with quotas and an administered exchange rate explains the trade far better than any account built around caravans and scenery. It also left an argument that has never really stopped. Whenever a state has treated tea as strategic rather than ordinary — the Qing on the frontier, the British with the Company’s monopoly, twentieth-century marketing boards with compulsory auction rules — the same questions recur about administered prices, licensed intermediaries and who captures the margin. The Song bureau is simply the earliest version that is documented well enough to be argued about.